Application of a hybrid algorithm to explicit estimation of bridge backwater

Bridges are one of essential infrastructures that are usually built on waterways and should be sustainability designed considering commonly flood events. Bridge afflux is an excessive rise of water surface during flood periods at upstream of a bridge which can be potentially harmful for bridge stability and safety. Predicting bridge backwater has been studied for decades which it implies the importance of this phenomenon in road and bridge constructions. In this regard, various models have been recommended based on experimental and field datasets for this purpose. In this paper, a new powerful hybrid optimization is utilized to develop a new explicit equation for estimation bridge afflux. A reliable database comprising both laboratory and field data was used to calibrate and testify the new equation. The performance of the new equation was compared with that of other explicit models available in the literature using five performance evaluation criteria. Based on four of the considered criteria, the recommended model outperforms other models in estimating bridge backwater for the test data. To be more specific, the new model improves the MSE and R2 values of the best available model by 39% and 14% for the test data. According to the obtained accuracy, the proposed equation in this study not only could estimate bridge afflux more accurately than other considered models but also may be utilized for backwater estimation within its applicable range in practice